RECORD
Multi-Dimensional Machine Translation Evaluation: Model Evaluation and Resource for Korean
Computation and Language · Submitted 19 Mar 2024
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim Overall, RemBERT emerges as the most promising model.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
Almost all frameworks for the manual or automatic evaluation of machine translation characterize the quality of an MT output with a single number. An exception is the Multidimensional Quality Metrics (MQM) framework which offers a fine-grained ontology of quality dimensions for scoring (such as style, fluency, accuracy, and terminology). Previous studies have demonstrated the feasibility of MQM annotation but there are, to our knowledge, no computational models that predict MQM scores for novel texts, due to a lack of resources. In this paper, we address these shortcomings by (a) providing a 1200-sentence MQM evaluation benchmark for the language pair English-Korean and (b) reframing MT evaluation as the multi-task problem of simultaneously predicting several MQM scores using SOTA language models, both in a reference-based MT evaluation setup and a reference-free quality estimation (QE) setup. We find that reference-free setup outperforms its counterpart in the style dimension while reference-based models retain an edge regarding accuracy. Overall, RemBERT emerges as the most promising model. Through our evaluation, we offer an insight into the translation quality in a more fine-grained, interpretable manner.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Almost all frameworks for the manual or automatic evaluation of machine translation characterize the quality of an MT output with a single number. | background 1.00 | 0.04 | 0.02 | asserted 0.99 |
| s1 | An exception is the Multidimensional Quality Metrics (MQM) framework which offers a fine-grained ontology of quality dimensions for scoring (such as style, fluency, accuracy, and terminology). | background 1.00 | 0.04 | 0.03 | asserted 1.00 |
| s2 | Previous studies have demonstrated the feasibility of MQM annotation but there are, to our knowledge, no computational models that predict MQM scores for novel texts, due to a lack of resources. | gap 0.99 | 0.07 | 0.03 | asserted 0.99 |
| s3 | In this paper, we address these shortcomings by (a) providing a 1200-sentence MQM evaluation benchmark for the language pair English-Korean and (b) reframing MT evaluation as the multi-task problem of simultaneously predicting several MQM scores using SOTA language models, both in a reference-based MT evaluation setup and a reference-free quality estimation (QE) setup. | contribution 0.98 | 0.90 | 0.08 | asserted 0.97 |
| s4 | We find that reference-free setup outperforms its counterpart in the style dimension while reference-based models retain an edge regarding accuracy. | result 1.00 | 0.96 | 0.16 | asserted 1.00 |
| s5 | Overall, RemBERT emerges as the most promising model. | result 0.99 | 0.90 | 0.47 | asserted 0.87 |
| s6 | Through our evaluation, we offer an insight into the translation quality in a more fine-grained, interpretable manner. | interpretation 0.97 | 0.87 | 0.19 | asserted 0.96 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
Almost all frameworks for the manual or automatic evaluation of machine translation characterize the quality of an MT output with a single number. An exception is the Multidimensional Quality Metrics (MQM) framework which offers a fine-grained ontology of quality dimensions for scoring (such as style, fluency, accuracy, and terminology). Previous studies have demonstrated the feasibility of MQM annotation but there are, to our knowledge, no computational models that predict MQM scores for novel texts, due to a lack of resources. In this paper, we address these shortcomings by (a) providing a 1200-sentence MQM evaluation benchmark for the language pair English-Korean and (b) reframing MT evaluation as the multi-task problem of simultaneously predicting several MQM scores using SOTA language models, both in a reference-based MT evaluation setup and a reference-free quality estimation (QE) setup. We find that reference-free setup outperforms its counterpart in the style dimension while reference-based models retain an edge regarding accuracy. Overall, RemBERT emerges as the most promising model. Through our evaluation, we offer an insight into the translation quality in a more fine-grained, interpretable manner.
Human review
How the literature uses it
What this paper does with the arXiv papers it cites
What it cites · 5 sentences without a confident label
Uses the method or tool of 2 papers
-
Concretely, we use Multilingual BERT base cased [ref] , XLM-RoBERTa large [ref] , and RemBERT [ref] (encoder-only) and Multilingual BART large-50 [ref] , Multilingual T5 large [ref] and M2M100 1.2B [ref] (encoder-decoder).
Rethinking embedding coupling in pre-trained language models · Base Model Selection · 0.96
-
Optimization is carried out with stochastic gradient descent, using the AdamW algorithm [ref] .
Decoupled Weight Decay Regularization · Experimental Setup · 1.00
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.